We shall first consider the utilization of molecular sieves in the subtraction method in more detail.
In the subtraction method molecular sieves have been successfully used as the selective
“reagent”; these sieves selectively adsorb and retain compounds that can pass through their openings owing to their geometry and size. The subtraction effect is based here on the physical properties of the molecules being analysed, rather than on their chemical properties. Molecular sieves were first used in the subtraction method by Brenner and Coates {34]. It was shown that C3-Cll n-alkanes are adsorbed quantitatively on a column containing molecular sieves 5A, whereas aromatic hydrocarbons, naphthenes and branchedchain aikanes pass through the column unaffected. Using this method, Brenner and Coates developed a technique for the quantitative determination of n-alkanes in reforming products. The total content of n-alkanes was determined from the difference in the peaks on the chromatogram without a reactor and with a reactor, i.e., by using the
165 conventional subtraction technique. The method was later used for analysing n-alkanes in petroleum fractions [35].
Brenner et al. [36] later extended this method to other classes of organic compounds, using it for the selective adsorption of alcohols, aldehydes, acids, esters and other com- pounds. It was established that molecular sieves 5A quantitatively adsorb propane, n-bu- tane, n-pentane, n-hexane, ethylene, propylene, hexene-2, methanol, ethanol, n-butanol, acetaldehyde, propionaldehyde, isovaleraldehyde and acetic and propionic acids. The following compounds pass through a column containing molecular sieves 5A: isobutane, isopentane, 2,3-dimethylbutane, benzene, toluene, xylene, cyclopentane, cyclohexane, isobutylene, 2-methylbutadiene-l,3, ethyl formate, ethyl acetate, ethyl propionate, acetone, methyl ethyl ketone, mesitelene oxide, methylene chloride, chloroform, iso- propanol, methlybutanol, diethyl and diisopropyl ethers, thiophene, carbon monoxide, methane, nitrogen dioxide, carbon sulphide, oxygen, nitrogen and nitromethane. Molecu- lar sieves 13X adsorb all of these compounds, with the exception of the gases (nitrogen, oxygen, carbon monoxide and methane). Molecular sieves 4 A adsorb only the lower members of the homologous series (methane, ethylene, propylene, methanol, ethanol and propanol).
Fig. 5.4 exemplifies the utilization of molecular sieves 5A in the subtraction method for the analysis of oxygencontaining compounds [36]. The chromatogram of the initial mixture (a) records five compounds: diisopropyl ether (l), propionaldehyde ( 2 ) , acetone (3), ethyl acetate (4) and ethanol (5). A chrcmatogram obtained with the use of a pre- column containing molecular sieves 5A (length 0.5 m, temperature 75OC) lacks the peaks of the compounds that are adsorbed (propionaldehyde and ethanol). Components of mixtures cannot be separated in this way by chenlical methods. Table 5.2 illustrates some examples of the molecular-sieve effect of zeolites. More detailed characteristics of zeo- lites are given in several books [37-40].
Molecular sieves, as physical reagents, are also used in the subtraction method for analysing rdatively high-boiling reagents. Thus, rapid methods with the use of molecu- lar sieves 5A [ 4 0 4 7 ] have been developed for the GC determination of n-alkanes in
?I oa
e
L c U
+ u n u
5 10 0 5 10 15 20
Time (min)
Fig. 5.4. Chromatograms of oxygen-containing compounds (a) without a pre-column and (b) with a pre-column filled with molecular sieves 5A. 2-m column packed with 30% Carbowax 1500 on Chromo- sorb; 0.5-m pre-column containing molecular sieves 5A. Column temperature, 75°C. For peaks, see text. From ref. 36.
166 TABLE 5.2
WOLtCULAR-SIEVE PROPERTIES OF ZEOLITES [ 1531 Type Compounds adsorbed
3A Helium, hydrogen, water
4 A Neon, argon, krypton, xenon, nitrogen, methane, ethane, ethylene, propane, acetylene, carbon dioxide, carbon disulphide, hydrogen sulphide, methanol, ammonia, methylamine, methyl bromide, methyl chloride
Propane, n-butane, n-heptane, tetradecane, ethyl chloride, ethyl bromide, ethanol, ethyl- amine, ethylene chloride, ethylene bromide, ethyl chloride, dimethylamine
Carbon tetrafluoride, carbon tetrachloride, chloroform, bromoform, isobutane and other isoalkanes, benzene, toluene, xylene, cyclohexane, thiophene, furan, pyridine, dioxane, naphthalene, quinoline, 6decyl-l,2,3,4-tetrahydronaphthalene
5 A
1 ox
13X 1,3,5-Triethylbenzene, 1,2,3,4,5,6,7,8,13,14,15,16dodecahydrochrysene, di-n-butylamhe kerosenes and gas oil. According to Barrel and Baumann [47] the accuracy of determi- nation of C7-Czo normal hydrocarbons is 2-3%.
An advantage of the subtraction method, which, however, can be used in only a limited number of instances, is the possibility of selective determination of the adsorbed compounds after their desorption. The adsorbed compounds are regenerated either by increasing the temperature or by the action of a new reagent, which destroys the com- pounds formed or displaces the components absorbed. In this instance the selectivity of determination increases considerably, and so do the reliability and sensitivity. An example of the utilization of this method is the determination of n-alkanes in a mixture with other hydrocarbons; n-alkanes absorbed at 350°C can be eluted at 550"C, whereupon their individual composition can be determined with temperature programming of the chro- matographic column, which is located after the subtraction reactor [48].
Molecular sieves are used mainly for separating n-alkanes from other types of hydro- carbons. The more limited utilization of molecular sieves in analysing compounds of other classes (for instance, oxygencontaining compounds) is due to the fact that com- pounds containing hydroxyl and carbonyl groups, when adsorbed on a moIecular sieve, may hinder the passage of other similar compounds through the sieve [49]. Unfortu- nately, the analytical properties of molecular sieves have been studied insufficiently well.
The potential of molecular sieves as highly specific adsorbents is utilized too inefficiently.
Detailed investigation into the analytical properties of molecular sieves wiU undoubtedly greatly expand the field of application of these unique adsorbents.
As has already been noted, in chromatographic practice molecular sieves are used particularly for the analysis of complex hydrocarbons, and chemical reagents are used together with the sieves. An interesting example of the utilization of subtraction methods was described by Rowan [SO], who developed a method of chemical absorption of par- ticular components of a complex mixture by using concentrated sulphuric acid (absorp- tion of olefins and aromatic hydrocarbons), mercury perchlorate (absorption of olefins) and molecular sieves (absorption of n-alkanes) as group absorbers. The proposed pro- cedure consisted in passing the initial sample through a chromatographic column, a
167 detector and a reactor to a cold trap. The products condensed in the trap were re-analysed on the column to establish the changes that had occurred in the test sample as a result of the typical reaction. Apart from removal reactions, Rowan used hydration of olefins and aromatic hydrocarbons and dehydration of naphthenes.
The analysis was carried out using a standard chromatograph with a slightly modified circuit (four-way valves, connecting lines, traps and reactors were introduced). The reac- tors (absorbers) were detachable. Chromatographic separation was carried out in a column containing dodecyl phthalate (length 200 cm) at 125"C, using hydrogen or helium as the carrier gas. The size of the test sample varied between 0.01 and 0.05ml. The freezing trap was a U-shaped tube (diameter 3 mm), filled with a packing (35% silicone oil IIC-200 on refractory brick) over a small section (5 cm) for better operation. An experimental check demonstrated the efficiency of this type of trap. Cooling was accomplished with liquid nitrogen, although dry-ice can also be used in many instances.
Selective absorption of olefins and aromatic hydrocarbons was carried out in an ab- sorber (diameter 6 mm) containing 2 ml of concentrated sulphuric acid on a layer of fibre-glass. The length of the packing layer was 43 cm. At the outlet of the absorber a layer of molecular sieve 4A was placed; this sieve adsorbs methane, water, ethane, acety- lene and propylene, but not propane and high-molecular-weight compounds. A few granules of Ascarite were also added. The reactor absorbs olefins, toluene and higher aromatic hydrocarbons at 54°C; benzene is absorbed incompletely.
Rowan [50] noted that removal of olefins with sulphuric acid has the disadvantage that small amounts of light hydrocarbons (3-6% of the olefin content), which cannot be separated from the hydrocarbons present in the sample, are released continuously for
1-2 h after the passage of the olefins through the absorber.
Selective removal of olefins was carried out in the absorber with mercury perchlorate, which was prepared in accordance with the recommendations of Coulson [51], who used it in mass spectrometric analysis. The adsorber was prepared by treatment of refractory brick (40-60 mesh) with 1 M mercury perchlorate solution-2 M perchloric acid (1 : 1, w/w) and subsequent drying at 110°C. During storage of the absorbent in a closed vessel, no decrease in its activity was observed. Rowan [50] used a layer of mercury perchlorate of height 25cm in a copper tube of 6mm diameter for absorption of olefins at 82 or 100°C. A layer (height 12.5cm) of molecular sieve 4A was used to adsorb the water released from the reactor. The method of selective absorption of unsaturated compounds was used successfully for analysing petrol.
Martin [52] proposed a combined method for determining aromatic, olefinic and saturated hydrocarbons in petrols. The method is based on functional group separation by chromatography and chemical absorption in a reactor containing mercury perchlorate.
First the test sample is separated on a column containing P,P'-thiodipropionitrile, from which a group of saturated olefin hydrocarbons is first eluted; after the detector these hydrocarbons pass through an absorber containing mercury perchlorate, then they enter the cold trap. The aromatic hydrocarbons are eluted from the column much later. Then the direction of the gas flow is reversed and the carrier gas passes successively through the absorber, the cold trap, the chromatographic column and the detector. After the aromatic hydrocarbons have been eluted from the column the trap is thawed and the saturated hydrocarbons are determined. An advantage of this method is the possibility of
SUBTRACTION METHOD 5 . 3
SORPTlON OF HYDROCARBONS OF DIFFERENT TYPES BY CHEMICAL SORBENTS Kcprintcd with permission from ref-. 9.
Hydrocarbon Sorption (7n)
2%HgSO, + Hg(CH,COO), 4% Ag,SO, + 95% H,SO, 80% 11, SO, 60% tl SO,
20% H, SO, 95% H , SO,
Methane 0 0 0 0 0 0
Ethane 0 0 0 0 0 0
Propane 0 0 0 0 0 0
n-Butane 0 0 0 0 0 0
n-Hexane 0 0 0 0 0 0
n-Octane 0 0 0 0 0 0
Cyclohexane 0 0 8 31 0 0
Ethylene 100 94 100 11 6 0
Propylene 100 100 100 100 67 0
Isobutylene 100 100 100 100 100 100
2-Pentene 100 67 100 100 88 0
2-Heptene 100 60 100 100 85 0
4-Methylc yclohexene 100 I 0 100 100 70 0
Benzene 5 46 100 94 32 13
Toluene* 0 22 100 100 33 0
p-Xylene* 100 0 100 I00 0 0
Acetylene 100 100 100 10 11 0
* Accuracy 10%. m c
W 4 P >
2 n Z s
m 4 0 s
U
IDENTIFICATION OF TEST MIXTURES 169 functional group analysis using a single sample. Olefinic hydrocarbons are determined by the difference between the total content of saturated and olefinic hydrocarbons and the content of saturated hydrocarbons.
A detailed investigation of various chemical absorbers was described by Innes et al. [9].
An appropriate chemical reagent was applied on diatomite (80-100 mesh) in the ratio of 1 ml of reagent solution per gram of diatomite. Reactors of length 10 cm and diameter 6.25 mm were used for selective absorption; the size of the test sample was 5 ml, con- centration 0.5%, and carrier gas flow-rate 20-50ml/min. Table 5.3 gives the results for various reagents.
Table 5.3 shows that a solution of silver sulphide-sulphuric acid absorbs acetylene and olefinic and aromatic hydrocarbons quantitatively, whereas a solution of mercury(I1) sulphide-sulphuric acid absorbs only acetylene and olefinic hydrocarbons. Concentrated sulphuric acid removes only defins (beginning with propylene) and aromatic hydrocar- bons; ethylene and acetylene are absorbed only partly.
To record the results of chromatographic separation and chemical absorption, Innes et al. 191 used a unique differential system consisting of two flame-ionization detectors and parallel absorbers; one detector recorded the concentration of hydrocarbons after the first absorber, and the other after the second absorber. By synchronizing the operation of the two detectors and obtaining a differential signal from them (electrical subtraction), it is possible to obtain a signal proportional to the concentration of the hydrocarbons absorbed (acetylene and olefinic hydrocarbon, etc.). Some versions of the use of ab- sorbers and detectors are given in Table 5.4. The method has been applied to the analysis of olefins and alkanes in exhaust gases and for analysing petroleum.
Albert [53] developed a GC method for determining the types of hydrocarbons (aro- matics, unsaturated, n-alkanes and isoalkanes) in mixtures of C5-Cll hydrocarbons. The method is based on the use of a selective liquid stationary phase, N,N-bis(2cyanethyl) formamide, from which aromatic hydrocarbons are eluted after the other compounds, molecular sieves, which retain n-alkanes selectively, and an absorber containing mercury perchlorate, in which unsaturated compounds are absorbed.
The absorber is filled with mercury perchlorate on Chromosorb to a height of 7.6 cm, anhydrous magnesium perchlorate to 5.1 cm, Ascarite to 5.1 cm and anhydrous mag- nesium perchlorate to 2.5cm. The analysis is performed in special chromatographic equipment consisting of a chromatographic column, an absorber and a trap for repeating the chromatographic analysis of some groups of hydrocarbons (isoalkanes, n-alkanes).
The n-alkanes adsorbed by the molecular sieves desorb into the trap on heating at 390- 400°C for 15 min. The duration of a complete analysis is 1.5 h. The method has been applied to petrol analysis.
Analysis of unsaturated compounds with the use of a reactor containing concentrated sulphuric acid has also been described [54,55]. For subtracting unsaturated compounds one can also use complexing agents, such as salts of silver [56,57] and mercury [58-601.
The use of the formation of involatile compounds in the GC analysis of hydrocarbons has also been described by other workers [61,62]. The subtraction method is also used successfully in the analysis of organic compounds of other classes.
The formation of involatile compounds has also been used for the selective removal of alcohols from a mixture of organic compounds. Ykeda et al. [63] used a reactor
TABLE 5.4
SOME VERSIONS OF UTILIZATION OF DIFFERENT SORBENTS
CCCK = 4c; Ag, SO, + 95% H 2 SO, ; PCCK = 2% HgSO, + 20% H 2 SO,. Reprinted with permission from ref. 9.
Absorbent Type of test hydrocarbons
Absorber 1 Absorber 2 Absorber 1 Absorber 2 Differential signal of detectors
~ ~ ~ ~~ ~ _ _ _ _ _ _ _ _ _ _ _
CCCK F'CCK Alkanes Alkanes + aromatic Aromatic hydrocarbons hydrocarbons
CCCK Glycerol Alkanes Au hydrocarbons Aromatic, acetylene and olefin hydrocarbons
PCCK Glycerol Alkanes t aromatic All hydrocarbons Olefin and acetylene hydro-
hydrocarbons carbons
(1 5 x 0.6 cm I.D.) containing 3% boric acid on Chromosorb P to remove alcohols as a result of the formation of involatile ethers of boric acid. Before use, the reactor was heated at 225°C. To establish the analytical potential of the method, pure compounds were separated on a column (305 x0.6cm I.D.) (containing 20% Carbowax 20M on Chromosorb P).
Fig. 5.5 depicts two chromatograms. The chromatogram on column A contains all of the components of the test mixture: tert.-butanol (1), n-butanol(2), n-cymol (a), linalool ( 3 ) , methanol (4), methyl phenylacetate (b), benzyl alcohol (5) and 2-acetylpyrrole (c).
The bottom chromatogram was obtained with column A, connected in series with a reac- tor containing boric acid. The components that did not contain a hydroxyl group (a,b,c) passed through the reactor. Peak 3 corresponds not to the initial alcohol (linalool), but to the product of its degradation (mercene). This method was used independently for separating terpene alcohols from terpene compounds of other classes [64] . For selective absorption of alcohols, use was made of a reactor filled with metaboric acid (0.1 g) on a
- I I I L V
5 1 0 15 20
Ttme (mln)
Fig. 5.5. Chromatograms of initial mixture (column A) and test mixture after separation and subtrac- tion icolumn B ) . For peaks, see test. I.'rorn rcf. 63.
support at 130-140°C. To convert the water formed on esterification, which interfered with the subsequent separation, a layer of a mixture of calcium hydride (1 g) with the support was placed in the reactor after the metaboric acid layer.
A systematic investigation of a reactor containing boric acid was carried out by Bier1 et al. [65] . The length of the packing layer was 150 mm; the packing contained 1 part of boric acid and 20 parts of sorbent with 5% Carbowax 20M on Anachrome ABC. Second- ary and primary alcohols are selectively retained by the packing, evidently in the form of boric ethers, whereas tertiary alcohols are usually dehydrated, forming olefins. However, the primary and secondary alcohols leave the reactor in the form of ethers very late and as very broad zones. Alcohols with a double bond in the a-position with respect to the secondary hydroxyl group were dehydrated and were not retained, like tertiary alcohols;
the same was true of alcohols with an ally1 bond. The phenol peaks broadened, their retention times increased (less than 2-fold) and salicylaldehyde was retained in the reactor with boric acid. Spatidy uninhibited carbonic acids were partially subtracted and their peaks were broadened, but their retention times did not increase more than 2-fold. Salicyl- aldehyde was retained quantitatively by the packing.
Osokin et al. [66] noted that in a microreactor containing 15% boric acid on diatomite some epoxy compounds are isomerized to aldehydes.
Prokopenko et al. [67] showed that tertiary alcohols can be removed completely in a reactor with an increased content of boric acid.
The formation of hydrogen bonds between the stationary phase and the test com- ponents can also be used for the selective retention of alcohols, acids and other com- ponents capable of forming hydrogen bonds. Ackman and Burgher [68] took advantage of the ability of high-boiling alcohols to form hydrogen bondswith a polyester stationary phase for determining non-alcoholic impurities in alcohols.
Together with group-specific reagents, it is often expedient to use specific reagents that react with one or two particular components. Reagents such as anhydrone, calcium chloride and phosphorus pentoxide (and also molecular sieves) are used, for instance, to absorb water, which hinders the chromatographic analysis of many compounds. Thus, in analysing aqueous solutions of hydrocarbons and 3-bromo-l , 1,2,2-tetrafluoropropane
[69], a reactor (452 x 0.6 cm I.D.) containing a mixture of phosphoric anhydride and refractory brick (weight ratio 9 : 1 ; brick fraction 60-80 mesh) was inserted before the chromatographic column. After absorption of water, chromatographic separation was carried out on a column (294 x 0.6 cm I.D.) packed with 20% silicone IIC-710 on refrac- tory brick. One volume of reactor packing can be used for the analysis of 50 samples of up to 0.1 ml each. The method was used for determining trace amounts of 3-bromo- 1,1,2,2-tetrafluoropropane.
Tosio et al. [70] suggested that in order to identify and determine acid components in mixtures with neutral compounds, a flow of carrier gas should be passed through a reac- tor (100 x 0.5 cm I.D.) packed with potassium hydroxide on a quartz powder (1 15 : 100) after the chromatographic separation of the initial mixture on an analytical column.
Selective absorption of the acid components takes place in this reactor. By comparing the chromatogram obtained with the analytical column and that obtained with an alkaline reactor, it is possible to identify and determine the acid and neutral components of the test mixture. As an example, results were presented of the analysis of small amounts of
phenols and cresols in a heavy oil from coal tar and showed that this method is also suitable for analysing compounds containing active hydrogen such as indene, fluorene, pyrrole, indole and carbazole, and also for identifying ketosteroids and oestrogens in steroid mixtures.
A reactor containing 20% Apiezon M on Celite 545 impregnated with a 5% solution of sodium hydroxide was used for removing fatty acids in the analysis of condensed milk extracts [71]. The use of sodium hydroxide on quartz for subtracting phenols has also been described [72]. As regards subtraction of acid products, the use of other basic reagents has also been described. The subtraction of organic compounds containing a carboxyl group with the use of a reactor filled with zinc oxide was proposed by Davison and Dutton [73]. Zinc oxide was used in the form of a powder deposited on the surface of the support or in the form of surface-oxidized zinc granules. The reactor containing zinc oxide (25 x 6.25 mm I.D.) was placed before the chromatographic column. Well oxidized zinc granules (ca. 30 mesh) or clean sand with an addition of 1-2% of zinc oxide was used as the packing.
Bierl et al. [65] showed that subtraction of acid products occurs in a short reactor (length 36mm) filled with a reagent consisting of a 1 : 10 (w/w) mixture of zinc oxide powder and polyethylene glycol adipate plus phosphoric acid. In this reactor the organic acids are subtracted completely, except for acids substituted in the a-position; for such acids 30-90% subtraction takes place. At the same time a broadening of the rear bound- ary of the zone is observed and also a 50% increase in the retention time. The change in the peak shape and retention time also helps to identify these acids. It should be noted that under these conditions partial subtraction of alcohols also occurs, the primary al- cohols are removed to the extent of 50% and the phenols, secondary and tertiary alcohols to the extent of less than 20%. The peak shape and retention time of alcohols and phenols remain unchanged.
Various reagents have been used for subtracting aldehydes and ketones. Thus, to re- move aldehydes and ketones from a mixture with alcohols and epoxides, Osokin et al.
[66] used a packing containing diatomite impregnated with an aqueous solution of sodium hydroxide (6%) and a solution of hydroxylammonium chloride in polyethylene glycol (20%). Aldehydes and ketones were converted into oximes and retained by the stationary phase, whereas alcohols and epoxides were not retained by this packing.
A layer of benzidine (20%) on Chromosorb P (36cm) removes aldehydes, most ketones and epoxides at 100-175°C; a-substituted ketones are retained only partly.
Retention of esters, ethers and alcohols is insignificant. According to Bierl et al. [65], subtraction of a compound by less than 45% of the amount introduced is not a reliable basis for its identification as an aldehyde, ketone or epoxide, as the reactor causes some retention and broadening of the absorption band even of non-reacting compounds in their passage through it.
The use of a reactor containing o-dianisidine makes it possible to subtract selectively most of the aldehydes, including &substituted compounds, but it does not remove ke- tones (except cyclohexanone), ethers, esters, phenols, olefins and hydrocarbons. Epoxides containing 12 or more carbon atoms in a molecule were subtracted partly or completely.
To identify epoxides, Bierl et al. [65] suggest the additional use of a column (reactor) containing phosphoric acid, whch effectively subtracts most of the epoxides.